the nature of inclination* tamar schapiro

Ethics 119 (January 2009): 229–256� 2009 by The University of Chicago. All rights reserved. 0014-1704/2009/11902-0003$10.00

229

ARTICLES

The Nature of Inclination*

Tamar Schapiro

According to the Kantian picture of action, our inclinations need notdetermine what we do. They influence us, but we have the capacity todecide, freely and rationally, whether or not to act on them.1 This pic-ture, though strongly associated with Kant’s theory, is not exclusivelyKantian. There are many rationalist theories according to which we havethe capacity to choose, in light of reason, how to act in the face ofinclination’s influence. In framing the question of this article, I amstarting from this general Kantian/rationalist picture of action, and I

* I am grateful to many colleagues for comments on earlier versions of this article.Special thanks to those who responded to presentations at the University of Toronto, JohnsHopkins University, Northwestern University, the University of California at Berkeley, Stan-ford University, Hampden-Sydney College, and the Universidad Nacional Autonoma deMexico. Thanks also to several anonymous referees and editors at Ethics for their extremelyhelpful comments.

1. I could have described Kant’s view another way. I could have said: “On the Kantianpicture of action, inclinations alone cannot determine what we do; in every action, wenecessarily choose whether or not to act on our inclinations.” That Kant believes this issuggested by his “incorporation thesis” (Immanuel Kant, Religion within the Boundaries ofMere Reason, in Religion and Rational Theology, ed. Allen Wood [Cambridge: CambridgeUniversity Press, 1996], 6:24; Henry Allison, Kant’s Theory of Freedom [Cambridge: Cam-bridge University Press, 1990], 40–41). This is the thesis that, in order to act on anyincentive (whether rational or sensuous), I must freely incorporate it into my maxim. Onthis view, the will is complicit in every action, including action from inclination. I believethat Kant did hold this thesis, and I believe that it can be defended, but it is highlycontroversial. Resistance to it comes largely from the fact that it tends to assimilate im-pulsive and unreflective actions to deliberate, reflective ones. A defense of the incorpo-ration thesis would have to address this problem. For my purposes in this article, I do notneed to assume the incorporation thesis. I only need to assume the weaker thesis that wecan take a step back from our inclinations and choose whether or not to act on them.This weaker thesis provides a less controversial starting point without compromising thesubstance of my argument. I am grateful to several readers, including Han van Weit-marchen and an anonymous reviewer at Ethics, for urging me to divorce my argumentfrom the stronger thesis.

230 Ethics January 2009

am attempting as far as possible to be neutral with respect to the variousways of fleshing out the concepts of freedom and reason that are as-sociated with it. My question is about the other concept, that of incli-nation. When I take a step back from my inclination for the purposeof deciding whether or not to act on it, what is it that I am steppingback from? What kind of thing is an inclination?

Pretheoretically, to be inclined is to be passively motivated. Incli-nations are in this sense “passions.” But the very notion of passive mo-tivation is paradoxical. Motivation is a form of self-movement. A stronggust of wind can move me, but it cannot motivate me. How, then, canthere be a form of motivation with respect to which I am passive? Whatis my inclination, such that it is both my motivation and something withrespect to which I am passive?

Before going further, let me make several clarificatory remarksabout the concept of inclination I am employing.2 While the notion ofinclination is closest to what contemporary philosophers call “desire,”it is narrower than that. ‘Desire’ is usually used to refer to a genericmotivational state in contrast to a generic cognitive state. This use ofdesire does not distinguish between those motivational states that exertan influence on the will and those that are states of the will. In Kant’sterminology, it does not distinguish between “pathological” motives,which have their source in our passive capacity for feeling, and “prac-tical” motives, which have their source in our active capacity for reason.3

In Thomas Nagel’s terminology, it fails to distinguish between “unmo-tivated” desires, which arise independently of deliberation, and “moti-vated” desires, which arise through practical deliberation. As I will beusing it, “inclination” refers to the first sort of motive in each pair.

Because it exerts an influence on the will, inclination is necessarilyaction-oriented. An inclination is, in the first instance, a motive to dosomething. I can have an inclination to drink a glass of gin, to run awayfrom an angry mob, to buy a new car, or to organize my closet. As such,inclination is not the same as a pro-attitude toward an object. I can havea pro-attitude toward you, but this itself does not amount to an incli-nation except insofar as it motivates me to do specific things, for ex-

2. I am not claiming that this is precisely Kant’s notion of inclination, but I also donot think that he used that concept consistently throughout his work (see n. 5). Giventhe essentials of his theory, I think that he would have recognized the concept of inclinationI describe here, and the questions relating to it, as valid. My hope is that Kantians andat least some non-Kantians will be able to do the same.

3. Immanuel Kant, Groundwork of the Metaphysics of Morals, in Practical Philosophy, ed.Allen Wood (Cambridge: Cambridge University Press, 1996), 4:399 and 413; Critique ofPractical Reason, in Wood, Practical Philosophy, 5:20; The Metaphysics of Morals, in Wood,Practical Philosophy, 6:399. I take Kant’s distinction to be equivalent to Thomas Nagel’sdistinction as it appears in his The Possibility of Altruism (Oxford: Clarendon, 1970), 29.

Schapiro The Nature of Inclination 231

ample, to call you or to make time to see you. I can have a pro-attitudetoward the state of affairs in which a meteor does not strike the earth,but unless I am in a position to do something about it, that attitudedoes not function as an inclination.4 Nor are inclinations necessarilypro-attitudes. They can involve aversion. Arguably, my inclination to runfrom an angry mob is best described as a kind of aversion rather thanas an appetite, a desire, or a pro-attitude.5

Nor is inclination obviously the same as emotion. Most emotionsare closely associated with, if not constituted by, inclinations. Fear, forexample, normally involves an inclination to avoid the feared object.But other emotions, such as grief, are less clearly tied to inclinations. Ileave it to others to clarify the precise relation between emotion andinclination as such.

In the foreground of much philosophical literature on desire arequestions about whether desires have belief-like content and about howdesires have satisfaction conditions, on analogy with truth conditions.These questions take their shape from an inquiry into the broad notionof desire, as a generic motivational state to be contrasted with a genericcognitive state. Because my interest is in the notion of inclination, mycentral concern is different. My aim is to show how inclination is similarto and different from volition. The resulting view may well have impli-cations for the other debates, but it is developed in response to a dif-ferent question.

In asking how inclination is similar to and different from volition,I am taking these concepts to have their home in the first-personalcontext of deliberation. I want to know how we have to conceive ofwanting and willing, given the respective roles we attribute to them inthe course of deliberation. As such, my aim is not to individuate mentalstates for the purpose of explaining and predicting phenomena third-personally. The aim is to give an account of the relation between in-clining and willing that could be illuminating to the agent who is en-gaged in both.

That agent has to be a creature of a certain kind, for the question

4. However, quite often hopes involving events beyond our control function as in-clinations to do things that we fantasize will affect the outcome. The obvious examplesare inclinations to act out of superstition, such as the inclination to wear a lucky shirt onthe day of the Big Game. More common is the fantasized, perhaps unconscious, attemptto control events beyond our control, such as events in the past, by obsessively rehearsingthings we could have done to make them turn out other than they did.

5. In the Groundwork, Kant uses the phrase “inclination or fear” as if inclination werea form of attraction to be contrasted with the aversion implicit in fear (Kant, Groundwork,4:398; 401n, and 440). I don’t think anything in his theory commits him to this, and inhis later Critique of Practical Reason he characterizes both hope and fear as inclinations(Kant, Critique of Practical Reason, 5:74 and 147).


I am asking only arises for a creature capable of distinguishing in de-liberation between her will and the inclination that influences it. Onewho is able to say “I want to” is already in such a position. To say “Iwant to” is already to make an implicit distinction between the will thatis influenced and the inclination that is influencing. An infant’s cry isperhaps the precursor of “I want.” But an infant is not yet sufficientlyreflective to distinguish, even implicitly, between her inclinations andher will.

Although the question I am asking is the agent’s question, it is notthe everyday question that arises in response to inclination. In everydaylife, the natural question I ask with respect to any given inclination is“Should I act on it?” My question is the somewhat more abstract oneof how I am to conceive of my inclinations as such in the course ofdeliberating about them. Are my inclinations simply psychological eventsthat I observe as I would internal weather? Or are they my doings, theearliest stages of actions I have already undertaken but have not yetmanifested in overt behavior? Is inclining something that happens tome, or is it something I do?

This way of framing the question allows me to map out the con-ceptual territory in a relatively straightforward way. In the first part ofthis article, I will argue that a theory of inclination has to navigatebetween two extremes, one of which assimilates inclination to an ex-ternal happening and the other of which assimilates it to an exerciseof will. I will then defend a middle way that has its roots in Plato andAristotle and that is arguably compatible with Kant.6 Inclination, I willargue, issues from a distinctive “part of the soul”; it is the exercise of asubpersonal capacity that is both agential and nonrational.

I. THE EXTREME ANTI-RATIONALIST VIEW OF INCLINATION

The temptation to assimilate inclination either to a happening or to adoing stems from ordinary intuitions about inclination, some of whichhighlight our passivity in relation to inclination and some of whichhighlight our active contribution to it. Let me start with the former.Intuitively, inclinations are passions. We are passive with respect to themin at least three senses. First, inclinations seem to come to us unbidden,spontaneously, as if from without. Inclinations bubble up in us, seizeus, wash over us, and assail us. Second, they are not directly responsive

6. My view echoes book 9 of Plato’s Republic, in which Plato describes the humanindividual as a hybrid of animal and human parts, whose “outer covering” makes himappear to be “a single creature, a human being” (Plato, Republic, in Plato: Complete Works,ed. John Cooper [Indianapolis: Hackett, 1997], 588d). It also provides one way of inter-preting Aristotle’s claim that the desiderative and emotional part of the soul is not strictlyrational but “partakes of reason in a sense” (Aristotle, Nicomachean Ethics [Indianapolis:Hackett, 1985], 1102b13–1103a).


to volition. We can have inclinations we wish we did not have, and wecan lack inclinations we wish we had. In neither case can we changeour inclinations simply by changing our minds. This does not mean wecannot change our inclinations at all. Exactly how we can is a furtherquestion, one that presupposes an answer to the question of this article.I will take it as a common intuition, though, that we cannot author ourinclinations in the same direct way that we author our actions. A third,related sense in which we are passive with respect to inclination is thatwe normally don’t hold ourselves responsible for having our inclinationsin the same way that we hold ourselves responsible for acting on them.Perhaps there is some attenuated sense in which we are responsible forhaving our inclinations, but I want to leave that question open. For mypurposes, the important intuition is simply that any responsibility wemight have for our inclinations is different, and less direct, than ourresponsibility for our actions.

These intuitions support the idea that inclination stems from asource external to reason or will. I will call this idea “the anti-rationalistinsight.” Now it is natural to take this insight as the basis for what I willcall “extreme anti-rationalism” with regard to inclination. Extreme anti-rationalism locates the motivational source of inclination in somethingwholly distinct from our agential capacities. Now, because I am startingfrom a broadly Kantian conception of our agential capacities, the viewI am calling extreme anti-rationalism is a form of dualism. The claimis that our inclinations are causally determined, whereas we freely authorour actions; inclination is the product of natural necessity, whereas ac-tions are products of reason. It is important, then, not to confuse ex-treme anti-rationalism about inclination with the more familiar sort ofHumeanism according to which desire and practical reason do not differdeeply in kind. As I will be using the term, “extreme anti-rationalism”refers to the dualist view, not the Humean view.

The attraction of extreme anti-rationalism is not hard to see. Theconception of inclinations as effects of brute causal processes seems tocapture the sense in which having an inclination to A involves feelinglike Aing or having an urge to A, where that feeling is something thatcan assail us independently of our volition. It also seems to account forthe ways in which we are passive with respect to our inclinations. If myinclinations are just effects of causal processes working through me, itis clear why I cannot be held responsible for having them and why Icannot change them simply by changing my mind. Despite this appeal,it is hard to find a well-developed exemplar of extreme anti-rationalism.Kant is sometimes caricatured as being one, but I believe his actual viewis in line with the one I will be defending later in this article. In general,those who start from a Kantian/rationalist conception of volition tendnot to flesh out their conceptions of inclination. When they do, it is


usually with the aim of defending a rationalist conception of inclinationagainst what is taken to be a crude dispositionalist picture.7 But thosewho develop and advocate the dispositionalist picture in any detail tendto be full-blown empiricists, and, hence, they don’t fit the model I amcalling extreme anti-rationalism.8 For this reason, my discussion of ex-treme anti-rationalism will be indirect. I will first approach disposition-alism through the eyes of a rationalist critic, and then I will ask whatthe dispositionalist might say in response to his worries.

In his well-known paper “Putting Rationality in Its Place,” WarrenQuinn argues against the view that desires are properly construed ascausal dispositions, unguided by rational thought. His stated target is atype of Humean, but it is a type of Humean who is like a dualist in thathe does distinguish between desires and reasons. Unlike Hume himself,this Humean accepts that there are such things as reasons for action.In addition, he claims that desires are sources of reasons, in the sensethat when we desire an end, we thereby have reason to promote thatend. But he also conceives of desires dispositionally. So, on the viewQuinn targets, desires are brute dispositions, unguided by rationalthought, that nevertheless “rationalize” actions by giving us reason totake the actions that satisfy them.

Quinn presents his argument as an internal critique of this position.He attempts to show that the Humean cannot jointly hold both thatdesires are dispositions and that desires are sources of reasons to act soas to satisfy them. But he then goes beyond an internal critique of theHumean position. After making the negative claim that the Humeanposition is internally incoherent, he goes on to argue for a positive claimabout how desires ought to be conceived. The argument is that, unlessdesires are conceived this way, they cannot rationalize action. Evidently,then, Quinn believes we have reasons independent of Humeanism tosave the thesis that desires rationalize action. Our task is to figure outwhat Quinn means by this thesis, why he thinks it is worth vindicating,and why he thinks the dispositional view cannot vindicate it.

The core of Quinn’s official argument is his example of a dispos-itional state characterized by nothing more than a tendency to turn onradios:

Suppose I am in a strange functional state that disposes me to turn

7. An exception is Ruth Chang’s “Can Desires Provide Reasons for Action?” in Reasonand Value: Themes from the Moral Philosophy of Joseph Raz, ed. R. Jay Wallace, Philip Pettit,Samuel Scheffler, and Michael Smith (Oxford: Oxford University Press, 2004), 56–90. Inthat paper, Chang takes aim at extreme rationalist conceptions of desire. But her con-trasting account of brute “feelings like it” is rather undeveloped.

8. For a succinct account of the dispositionalist conception of desire, see MichaelSmith, The Moral Problem (Oxford: Blackwell, 1994), 113–16.


on radios that I see to be turned off. Given the perception that aradio in my vicinity is off, I try, all other things being equal, to getit turned on. Does this state rationalize my choices? Told nothingmore than this, one may certainly doubt that it does. But in thecase I am imagining, this is all there is to the state. I do not turnthe radios on in order to hear music or get news. It is not that Ihave an inordinate appetite for entertainment or information. In-deed, I do not turn them on in order to hear anything. My dispo-sition is, I am supposing, basic rather than instrumental. In thisrespect it is like the much more familiar basic dispositions to dophilosophy or listen to music.9

Quinn argues that this dispositional state does not rationalize the actionsit motivates, in the sense that it fails to generate “even a prima faciereason to turn on radios” and “does not make the act sensible.”10 Whathe seems to have in mind here is a very modest notion of rationalization.The point is not that the desire fails to justify the action from someobjective point of view, but rather that it does not give the agent aperspective from which to see the action as worth doing, even in aminimal sense.

Quinn’s positive claim is that the only way to show that desiresrationalize actions is by abandoning the dispositional account of desirein favor of what I will call (following Talbot Brewer) the “evaluativeoutlook” conception of inclination.11 On this view, desires are evaluativeperspectives on the world, perspectives that are shaped by the agent’simplicit evaluations of objects and actions as good or bad. As Quinndefines it, the notion of evaluation being employed here is quite broad.It encompasses not only explicit judgments of good and bad in moraland prudential senses but also judgments that the object in question isattractive or pleasant, which, Quinn argues, are evaluative rather thanstraightforwardly empirical judgments.12 The claim is that desire canrationalize action only if it involves my seeing or thinking of the actionas good in this broad sense.

Now there are several weaknesses in this argument. First, if Quinn’sdiagnosis is correct, the sense of “rationalize” being employed in theargument is so weak as to not be distinctive of the kind of Humeanismhe has in mind. The sense in which the radio man’s desire fails to

9. Warren Quinn, “Putting Rationality in Its Place,” in Virtues and Reasons: PhilippaFoot and Moral Theory, ed. Rosalind Hursthouse, Gavin Lawrence, and Warren Quinn(Oxford: Clarendon, 1995), 189–90.

10. Ibid., 190.11. Talbot Brewer, “Three Dogmas of Desire,” in Values and Virtues: Aristotelianism in

Contemporary Ethics, ed. Timothy Chappell (Oxford: Oxford University Press, 2006), 257–84.As will become clear later, Thomas Scanlon shares this conception.

12. Quinn, “Putting Rationality in Its Place,” 197.


rationalize his action is simply that it fails to give him a perspective fromwhich to see the act of turning on radios as having any point. Now evena Kantian could agree with the claim that a desire (or, in my terminology,an inclination) has to rationalize action in this sense. This need not beinterpreted as equivalent to the claim that desiring to A gives the agenta reason to A, or even a partial justification for Aing , if by that is meantthat it gives him a reason or a justification from some more objectivepoint of view. If the Humean is committed to that stronger claim, thenQuinn’s solution to his problem misses its target.

Second, it isn’t clear that Quinn’s solution is unavailable to theHumean. The dispositionalist is not committed to the view that just anydisposition to act counts as a desire. Presumably the dispositions thatcount as desires are those that fulfill a set of narrower conditions. Hence,the dispositionalist could simply deny that the radio man’s disposition,as Quinn has described it, is the kind of state that counts as a desire.David Copp and David Sobel mount this sort of defense in response toQuinn on behalf of the dispositionalist.13 They maintain that the ten-dency displayed by radio man is too “thin” to count as anything otherthan a bare impulse, like the impulse to blink. Desires, they claim, arerelatively “thick” dispositions constituted not only by tendencies to actbut also by tendencies to entertain certain thoughts, including thethought that “turning on radios would be pleasant.” More might needto be said to distinguish a desire to turn on radios from, say, a psycho-logical compulsion to turn on radios, but the point is simply that thesophisticated dispositionalist in principle has resources to distinguishbetween mere impulses, psychological compulsions, and desires.

This defense is sound as far as it goes. But I think that somethingmore complicated is going on in the dialogue here. The dispositionalistis attempting to give necessary and sufficient conditions for attributinga desire to an agent from the third-person point of view. Copp and Sobelwrite: “We agree . . . that motivation by desire does not seem to theagent to be motivation by a disposition nor by a complex set of dis-positions.”14 But I think Quinn is making a point about how we have toconceive of our desires from the deliberative point of view, from thepoint of view of the agent who is influenced by them. Quinn writes:“That I am set up to head in a certain way cannot by itself rationalizemy will’s going along with the set-up. For that I need the thought thatthe direction in which I am psychologically pointed leads to somethinggood (either in act or result), or takes me away from something bad.”15

13. David Copp and David Sobel, “Desires, Motives, and Reasons: Scanlon’s Ratio-nalistic Moral Psychology,” Social Theory and Practice 28 (2002): 243–76.

14. Ibid., 269.15. Quinn, “Putting Rationality in Its Place,” 195.


Quinn’s challenge is posed from the standpoint of one who is underthe influence of inclination and who is sufficiently reflective so as todistinguish between his inclination and his capacity for choice. If thisis right, then the dispositionalist and Quinn may simply be talking pastone another. The dispositionalist grants that desires don’t appear to bedispositions from the agent’s point of view. He would also grant thatwhat Quinn calls “my will” doesn’t appear to be a disposition from theagent’s point of view, but he would nevertheless claim that it is possibleto give a dispositional account of it.16 Quinn does not explicitly argueagainst this methodology, but the methodology he adopts is different.His claim is that, when I am influenced by desire from the deliberativestandpoint, I cannot simply regard it as a dispositional tendency. Whynot?

It is one thing to claim that, from the deliberative standpoint, it isimpossible to conceive of my will as a mere disposition; it is another todeny that I can conceive of my inclination that way. I can’t conceive ofmy will as a mere disposition when I am engaged in deliberation becauseit is a presupposition of that activity that my actions are up to me andthat I can determine myself freely, according to what I see as the bestreasons. This presupposition holds regardless of whatever tendencies Ihave displayed in the past or will display to a third-person observer whenI act in this case. But insofar as my inclination presents itself to me, inthe deliberative standpoint, as something independent of my will, whycan’t I conceive of it as a mere disposition?

According to Quinn, were I to conceive of my desires as dispositions,they would no more rationalize my actions than would a disposition tosneeze rationalize my sneezing.17 By this, I take it Quinn means that mydesires would not give me a perspective from which to see the point ofacting on them. But why isn’t this argument diffused by thickening therelevant disposition so as to include evaluative thoughts? The dispositionthat counts as ‘desiring to eat some coffee ice cream’ no doubt differsfrom the disposition to sneeze when allergens are present because theformer motivates by way of my thoughts while the latter doesn’t. So whycan’t I, from the deliberative point of view, regard my desire to eatcoffee ice cream as a complex disposition involving certain patterns ofevaluative thought?

Quinn does not explicitly address this question, but I think he needsto answer it, and I will try to offer an answer on his behalf. Suppose I

16. Copp and Sobel argue that the Humean can give a dispositionalist account ofrationality as well as of desire (Copp and Sobel, “Desires, Motives, and Reasons,” 251).

17. Quinn, “Putting Rationality in Its Place,” 198.


feel like eating some coffee ice cream.18 And suppose I understand thatdesires are complex dispositions. I understand that, in having this desire,I am displaying a tendency, for example, to be assailed by thoughts ofthe pleasure of eating coffee ice cream and to experience certain phys-iological effects in response to these thoughts. Now let’s suppose thatI take up the deliberative point of view while still regarding my desireas a disposition of this sort. What, if anything, goes wrong?

If I really am viewing my desire as a disposition, then I am notoccupying it as an evaluative outlook. I am not deliberating aboutwhether to eat some coffee ice cream from the perspective of one whosees doing so as pleasant. Rather, I am outside that perspective, observingmyself having that pattern of thought. As such, it is not even clear whatthe deliberative question is. I am simply observing a feature of my cir-cumstance as I might observe any other feature of my circumstance.Perhaps I could ask whether having this disposition counts as evidenceof any other fact that might be relevant to determining what reasonsapply to me. It might be the case that, when I display this sort of dis-position, my hormones tend to be imbalanced. Or it might be the casethat I tend to crave coffee ice cream when I am subconsciously tryingto procrastinate working on a project. It also might be the case that myhaving this disposition is simply evidence that I would be likely to enjoythe experience of eating some. Some of these facts might be groundsfor concluding that it would make sense for me to eat coffee ice cream,while others might be grounds for concluding that it wouldn’t. By thesame token, some of these facts might be grounds for concluding thatit would make sense for me to perform some other action having noth-ing to do with eating coffee ice cream, like taking a hormone pill. Thepoint is that, on this picture, there is no necessary connection betweenmy wanting to eat some coffee ice cream and my having a perspectivefrom which it makes sense for me to do so. More fundamentally, thereis no necessary connection between my wanting to eat some coffee icecream and my raising the question “Should I eat some?”

While Quinn’s aim is to vindicate the thesis that acting on my desiresnecessarily makes some kind of sense to me, he actually vindicates a

18. I take this example from Thomas Scanlon, What We Owe to Each Other (Cambridge,MA: Harvard University Press, 1998), 44–45. As will become clear in the next section,Scanlon’s official view is extreme rationalism, not extreme anti-rationalism. But this is onepoint in his text where he imagines the agent taking an external perspective on his desires,regarding them as facts among others to be taken into account in determining what reasonshe has. Copp and Sobel share my discomfort with Scanlon’s description of the agent’spoint of view in this example, but their objection is rather undeveloped: “When I say ‘Iwould like some ice cream’ and head to the freezer, I am expressing my desire and actingon it, not reporting a hypothesis about enjoyment that I have formed on the basis ofnoticing my desire” (Copp and Sobel, “Desires, Motives, and Reasons,” 271).


more fundamental thesis. He vindicates the thesis that desires are thekinds of things we can act on. If my desire is merely an observed fact,a feature of my practical circumstance, I can take it into account indeciding what to do. But I cannot act on it as I would a proposal or ademand or a suggestion. Facts and proposals play distinct and mutuallyexclusive roles in deliberation. A fact contributes to determining whatmy circumstances are; a proposal attempts to tell me what to do givenmy circumstances. Our most fundamental notions about desire’s influ-ence on the will depend on the idea that it plays the role of a proposal.It is not just that we think of desires as things we can act on. We alsothink of desires as things that we can satisfy by acting on them. But thevery notion of satisfying our desires makes no sense if we regard ourdesires simply as events. Neither does the idea that desire can conflictwith reason. Reason cannot govern events, nor can it come into conflictwith them. The source of the intuition that desire and reason can con-flict comes from taking desires as proposals about what do to. If ourdesires propose or demand or suggest that we do certain things, thenthey at least appear to be in the same business as practical reason,namely, that of governing our actions.

The implications of dispositionalism, when regarded as a thesisabout how we should conceive of our desires from the deliberativeperspective, are thus quite radical. What we would have to give up isnot simply the thesis that desires make sense of actions based on thembut also the more fundamental conception of desires as proposals thatwe can act on, satisfy, and side with. In assimilating desires to happen-ings, dispositionalism overlooks the fact that desires don’t simply moveus; they bid us to act.

II. THE EXTREME RATIONALIST VIEW OF INCLINATION

These considerations provide the deepest reasons to reject an extremeanti-rationalist conception of inclination. The question is how far to goin the other direction. As I have laid out the continuum, the extremeopposite view would be a form of monism that assimilates inclinationto an exercise of will or practical reason. Opponents of anti-rationalismtend to move in this direction by emphasizing the similarity betweenthe form of motivation involved in inclination and that involved involition. As we have seen, Quinn holds that desire, like volition, is nec-essarily guided by the agent’s evaluative outlook. Thomas Scanlon con-tinues this line of thought in his book, What We Owe to Each Other.19

19. Scanlon, What We Owe, 37–41. My focus in this article will be limited to the shortsection of Scanlon’s book entitled “Reasons and Desires: Motivation.” I will not considerScanlon’s further views about the justificatory role of desire because that would take mebeyond the scope of my primary question.


Scanlon’s view is a variation on Quinn’s in that it takes the notion of areason for action, rather than that of goodness, as the primary normativenotion. On Scanlon’s view, having a desire to A essentially involves takingcertain considerations as reasons to A. Having a desire to buy a newcomputer, for example, involves taking certain considerations (e.g., thatnew computers look good, that they have useful features) as reasons tobuy a new computer.20 Thirst, Scanlon claims, involves seeing certainconsiderations (e.g., that my throat is uncomfortably dry, that a cooldrink would relieve that discomfort) as reasons to drink.21

Scanlon intends this account to be true of all desires, whetherpractical or pathological, motivated or unmotivated. Unlike Hume, whoargued that reason alone cannot motivate, Scanlon argues, in effect,that passion alone cannot motivate. His claim is that the motivationalforce behind all action comes from the agent’s taking-something-as-a-reason-to-act. Now Kantians and some rationalists might agree with thisas an account of the motivational force of those desires that are gen-erated through practical reasoning. But it is a further question whetherthey would agree with it as an account of the motivational force of thedesires that purport to bubble up spontaneously. The question, in Kant’sterms, is whether Scanlon’s view assimilates all pathological motives topractical ones.

Scanlon is aware that the challenge for his theory is to show that itaccounts for what we intuitively think of as the spontaneous form of desire.He confronts one aspect of this worry in the following passage:

Having what is generally called a desire involves having a tendencyto see something as a reason. Even if this is true, however, this isnot all that desire involves. Having a desire to do something (suchas to drink a glass of water) is not just a matter of seeing somethinggood about it. I might see something good about drinking a glassof foul-tasting medicine, but would not therefore be said to havea desire to do so, and I can even see that something would bepleasant without, in the normal sense, feeling a desire to do it.22

The notion of desire that Scanlon is referring to here is the narrow notion.If it were the broad notion of desire, he would have to be interpreted asmaking the Humean point that reason alone cannot motivate. But Scan-lon does not believe this. He would not deny that seeing something goodabout drinking a glass of foul-tasting medicine is sufficient to motivateaction. The question is how that kind of motive is related to the kind ofdesire that at least purports to have a life independent of the will.

Scanlon’s response is to account for what appears to be a difference

20. Ibid., 43.21. Ibid., 38.22. Ibid., 39. Compare Quinn, “Putting Rationality in Its Place,” 203–4.


in kind in terms of a difference in degree. Instead of saying that path-ological desires differ from practical desires in being motivated by anoncognitive rather than (or in addition to) a cognitive attitude, heclaims that they differ in being motivated by a more rather than a less“insistent” evaluative outlook: “Reflection on the differences betweenthese cases leads me to what I will call the idea of desire in the directed-attention sense. A person has a desire in the directed-attention sensethat P if the thought of P keeps occurring to him or her in a favorablelight, that is to say, if the person’s attention is directed insistently towardconsiderations that present themselves as counting in favor of P.”23

According to Scanlon, desire in the pathological sense consists inbeing motivated insistently by the thought of certain considerations asreasons, as opposed to simply judging or seeing certain considerationsas reasons in a noninsistent way. Indeed, Scanlon maintains, it is a mis-take to take the felt difference between pathological and practical mo-tivation to mark a difference in kind:

We should not take “desires” to be a special source of motivation,independent of our seeing things as reasons . . . when a persondoes have desire in the directed-attention sense and acts accord-ingly, what supplies the motive for this action is the agent’s per-ception of some consideration as a reason, not some additionalelement of “desire.” Desire in the directed-attention sense char-acterizes an important form of variability in the motivational efficacyof reasons, but it does this by describing one way in which thethought of something as a reason can present itself rather than byidentifying a motivating factor that is independent of such athought.24

It is in virtue of this claim that I consider Scanlon’s position to be aversion of what I want to call “extreme rationalism.”25 Extreme rationalismstarts from the main rationalist insight, namely, that inclination engagesus as agents. It then takes this insight to imply that inclination engagesus as full-fledged rational agents. In particular, extreme rationalism de-nies the Platonic and Aristotelian view that there are agential parts ofthe soul in any philosophically deep sense. It denies that there aredistinctively passive and active motivational capacities, each making adifferent contribution to action. Instead, extreme rationalism holds thatthe soul is unitary, in the sense that agency involves the exercise of onerational capacity. This capacity can generate the appearance that it is

23. Scanlon, What We Owe, 39.24. Ibid., 40.25. Quinn may or may not embrace extreme rationalism. His position, as presented

in the paper I have been discussing, is not sufficiently developed to distinguish clearlybetween extreme rationalism and other views.


limited by another, but this only because it can be exercised with varyingdegrees of “insistence,” accuracy, clarity, and the like.

Can extreme rationalism account for the intuitions that drive anti-rationalism, intuitions about our passivity with respect to pathologicaldesire? Another place Scanlon attempts to show that it can is in dis-cussion of recalcitrant desires (“irrational thoughts”) and akrasia. Thesephenomena seem to involve conflict between distinct motivationalsources. How are we to account for them on a theory in which the soulis unitary? One way is to claim that what appears to be conflict is infact vascillation: when I feel tempted to reach for the cigarette againstmy better judgment, I am in fact vascillating between the judgment thatI have reason to smoke the cigarette and the judgment that I do nothave reason to smoke the cigarette.

Interestingly, Scanlon does not take this route. Instead, he refineshis view, introducing a new distinction in kind, one that is purportedlyshallower than the distinction he denies. Instead of simply claiming thatwe have one capacity to “take” considerations as reasons, Scanlon claimswe have one capacity that can be exercised in two distinct ways. Hewrites: “Being a [rational] creature involves not only the capacity tomake certain judgments and to be consistent about them, but also theability to see certain considerations as reasons and to think of and seeas reasons those things one has previously judged to be such.”26 Whendesires seem to conflict with reason, as in cases of temptation and ak-rasia, what has happened is that our capacity to “see” reasons has becomedissociated from our capacity to “judge” reasons. We judge that X is infact not a reason to A, yet we persist in seeing X as if it were a reasonto A. Scanlon writes: “Even if, for example, I have convinced myself thatI should not be influenced by the approval or disapproval of a certaingroup, I may find myself wondering anxiously what they would thinkof something I am considering doing. When these thoughts occur, Imay dismiss them immediately. Nonetheless, insofar as they involve (per-haps only momentarily) seeing something as a reason that I judge notto be one, they are instances of irrationality.”27

Now, I should note that Scanlon is not entirely systematic in hissubsequent use of the terms ‘seeing’ and ‘judging’. Later he writes thata desire in the directed-attention sense to buy a new computer “involvesa tendency to judge that I have reason to buy a new computer.”28 Theslip raises the question of how much philosophical weight he actuallyputs on the distinction between seeing and judging. Still, Scanlon moreoften describes the experience of desire as one in which a certain con-

26. Scanlon, What We Owe, 40.27. Ibid., 40.28. Ibid., 43.


sideration simply “presents itself” to one’s mind as a consideration infavor of doing an action.29 And he writes: “What I am claiming . . . isnot that all desires arise from prior judgments but rather that havingwhat is generally called a desire involves having a tendency to see some-thing as a reason.”30

The question now is whether, in positing a difference in kind be-tween seeing and judging, Scanlon has in fact compromised his claimthat desire and reason are not distinct motivational capacities. In makingthis distinction, Scanlon draws an implicit analogy between desire andperception.31 The analogy is powerful because the relation between per-ception and theoretical judgment is indeed in many respects analogousto the relation between desire and practical judgment. Perception pur-ports to give us reasons to believe, and desire purports to give us reasonsto act. Perception assails us spontaneously, as if from without, as doesdesire. Perception can conflict with and be recalcitrant to theoreticaljudgment, as when a stick in water appears bent even though we judgeit to be straight. Similarly, desire can conflict with and be recalcitrantto practical judgment, as when a cigarette appears as good-to-smoke,even though we judge that it isn’t. Finally, in cases of conflict betweenperception and theoretical judgment, we think we ought to side withjudgment. The same is true in cases of conflict between desire andpractical judgment.

Because of these analogous relations, the claim that desiring is akind of seeing has intuitive pull. But the explanatory value of the analogyis limited. Absent an independent theory of perception, it does notexplain why perception exhibits the distinctive features that make itanalogous to desire. What explains the fact that our perceptions canseem to assail us spontaneously and that they can be recalcitrant toreason? Why are we not responsible for our perceptions in the sameway that we are responsible for our judgments? Is it because perceptionand theoretical judgment issue from distinct cognitive capacities? If so,in what sense are the capacities distinct? Is the distinction between

29. “Sometimes the pleasure of eating coffee ice cream keeps coming to mind, pre-senting itself as a reason for getting some now” and “ I often choose one route ratherthan another . . . just because it is the alternative that presents itself as attractive at thetime” (ibid., 44 and 47–48).

30. Ibid., 39.31. This analogy comes up elsewhere in recent literature on desire and practical

reason. See, e.g., Dennis Stampe, “The Authority of Desire,” Philosophical Review 96 (1987):335–81; R. Jay Wallace, “Addiction as Defect of the Will: Some Philosophical Reflections,”Law and Philosophy 18 (1999): 621–54; Cheshire Calhoun, “Cognitive Emotions?” in WhatIs an Emotion? Classic Readings in Philosophical Psychology, ed. Robert C. Solomon (Oxford:Oxford University Press, 2003), 236–47; Christine Tappolet, “Emotions and the Intelligi-bility of Akratic Action,” in Weakness of Will and Practical Irrationality, ed. Sarah Stroud andChristine Tappolet (Oxford: Clarendon, 2003), 97–121.


perception and judgment a difference in kind or just one of degree(e.g., of insistence)?

I am not suggesting that the perceptual analogy is on the wrongtrack. My point is simply that the very questions at stake in Scanlon’sdiscussion of desire can be raised with respect to perception. As such,the analogy alone does not give us insight into the nature of desire. Allwe have is a set of parallel intuitions. Indeed, we could just as easily runthe analogy the other way. We could say that perception is a form of,or is analogous to, inclination. Perception is an inclination to believe,just as desire is an inclination to act. Without an independent theoryof inclination, this wouldn’t explain anything about perception.

To what extent, then, does Scanlon go beyond the appeal to anal-ogy? We have seen that he holds (1) that the content of the mentalstate he calls “seeing” reasons differs from what he calls “judging” rea-sons only in the degree of insistence with which the considerations infavor of Aing “present themselves” to us and (2) that seeing and judgingare fundamentally exercises of the same motivational capacity. In thenext section, I will explain why I disagree with claim 1. For now I justwant to point out that if Scanlon holds to claim 2, he needs to say moreabout how this is consistent with the idea that we are not responsible forhaving our desires in the same way that we are responsible for acting onthem. If I am responsible for acting on my desires, this is presumablybecause my actions have their source in the seat of my agential authority,something identifiable with my proper or active self (whether we call this“will” or “reason” or “choice”). If, as Scanlon claims, my desires have thesame source, then it would seem I should likewise regard them as ex-ercises of my agential authority and that I should hold myself responsiblesimply for having them. Scanlon would, of course, reject this conclusion,but he needs to say more about why his theory does not commit himto it.

He does say at one point that what seeing and judging have incommon is that the “motivational force of these states lies in a tendencyto see some consideration as a reason.”32 But we already know that bothseeing and judging take reasons as their objects. The question is howthe activities differ such that we are more directly responsible for thelatter. Scanlon comes closest to addressing the responsibility issue whenhe writes that “neither of these tendencies [to see reasons and to judgereasons] is wholly under the control of a normal person.”33 If he wereto add that seeing reasons is less in our control than judging reasonsand that responsibility depends upon degrees of control, then perhapshe could begin to explain why we are less responsible for our desires

32. Scanlon, What We Owe, 40.33. Ibid., 40.


than our actions. But he would have to explain why seeing reasons isless in our control than judging reasons, given that it is an exercise ofthe very same capacity.

I mentioned earlier that the standard way for an extreme rationalistto account for recalcitrant desires and akrasia is to characterize it as avacillation rather than as a conflict between distinct motivationalsources. I also noted that Scanlon does not take this route, appealinginstead to a distinction between “seeing” and “judging” reasons. Thequestion is whether this distinction, were he to spell it out in detail,would compromise his claim that desire is not a distinct motivationalsource. If it would, then Scanlon would have to choose between hisperceptual analogy account of recalcitrance and akrasia and his officialcommitment to a unitary picture of motivation (what I am calling “ex-treme rationalism”).

What would his picture look like were he to hold on to extremerationalism? The most natural and consistent way to fill out the extremerationalist view is to conceive of desire as a sort of hasty, unreliable actof judgment.34 The difference between desire and reason, on this view,is a difference in degree. Reasoning and desiring are the same activity,but what we call ‘reasoning’ is a more deliberate and reliable exerciseof this activity than ‘desiring’. One strength of this view over a non-cognitivist view of desire is that it would explain how desire and reasoncan interact, how desires can present claims suitable for direct evaluationon the basis of reason. For, on this view, to reflect on one’s desires issimply to double-check the hasty reasoning that led to the conclusionsimplicit in them.

But this rather straightforward version of extreme rationalism runsafoul of the worry about responsibility. There is no reason to think weshould be less responsible for exercising our reason hastily than we arefor exercising our reason carefully. That the judgments involved in ourinclinations are hasty and unreliable might give us reason to reflect onthem, to double-check them, as it were, but it does not give us an excusefor having made them.

Now, I am not suggesting that any extreme rationalist would will-ingly embrace this implication. However, it is not clear how he couldescape it. The claim is that desiring and reasoning are at bottom ex-ercises of the same capacity. Practical reason, insofar as it is the sourceof action, is the seat of agential authority. Hence, desires, too, mustissue from the seat of agential authority. Extreme rationalism thus as-

34. Robert Solomon holds something like this view of emotions generally, namely,that they are modes of thought designed for “emergency” situations. Emotions, he writes,are “urgent judgments” (Robert Solomon, “Emotions and Choice,” in Solomon, What Isan Emotion? 224–35).


similates desire to volition, and in doing so, it overlooks the fact thatdesires are not attributable to us in the same way that actions are.

III. INCLINATION AS ANIMAL ACTION

There are good philosophical reasons, then, to avoid both anti-ratio-nalism and rationalism in their extreme forms. We want to avoid assim-ilating inclination to a happening, and we want to avoid assimilating itto a doing. The former approach holds that the distinction betweeninclination and will marks a difference in kind; the latter holds that itmarks a difference in degree. Is there a middle way?

In this section I will argue that the way to avoid both extremeswhile preserving their respective insights is to distinguish between twoagential capacities that jointly characterize us as human agents. One isa capacity to demand and offer justifications to ourselves and so to takeconsiderations as reasons. The other is a more primitive capacity to seeobjects as calling for certain responses, independent of any justification.It is the latter capacity, I claim, that accounts for the motivational forceof inclination.

Let me start by raising an objection to Scanlon’s account of contentof thoughts that motivate us when we are under the influence of incli-nation. Scanlon is right that his account of the motivating thoughtinvolved in inclination—the thought that X is a reason to A—impliesextreme rationalism. But there are independent reasons to reject thataccount. I want to argue that Scanlon’s account intellectualizes incli-nation, freighting it with a layer of reflection that it does not have,simply qua inclination. Suppose I am terribly thirsty. I have been hikingin the California hills on a hot summer day, and I have run out of water.My throat is painfully dry, and I am aching for a drink. I have lost theability to enjoy my surroundings, and my mind is preoccupied withfiguring out when and how I will be able to quench my thirst. Scanlonwould claim that, in this situation, I am insistently seeing the drynessin my throat as reason to drink water and this is what is motivating me.But I contend that a more primitively normative thought could sufficeto account for the content and motivational force of inclination. It isnot that I am seeing the dryness in my throat as a reason to drink waterbut rather that I am seeing water as to-be-drunk.

This claim differs from Scanlon’s in two respects. First, the contentof this latter consciousness is not a statement of a purportedly normativefact.35 It is not the thought that “water is to be drunk by someone,” northat “water is to be drunk by me.” It is more like an imperative—“Drink!”

35. My view here is in sympathy with John Deigh’s criticisms of certain cognitivisttheories of emotion (John Deigh, “Cognitivism in the Theory of the Emotions,” Ethics,104 [1994]: 824–54).


or “Drink this!” or “Drink water!”36 My thirst involves my seeing waterin an imperatival mode, seeing it as “calling for” drinking or as “callingupon me” to drink it. Second, the salient normative thought is not thatof a reason. It is more like a thought of obligation, law, or practicalnecessity. And yet it does not go so far as to involve this concept. Rather,it is an unreflective experience of practical necessity, an experience thatdoes not involve consciousness of practical necessity as something thateither has or requires justification. The thought is not “Drink! because. . .”; it is simply “Drink!”37

An extreme rationalist would be skeptical of the suggestion that this“calling for” relation need not involve justificatory thought. For, if I seewater as to-be-drunk, this must be in virtue of something, namely, in virtueof the thirst-quenching features of water to which my attention is insis-tently drawn. Shouldn’t I say, then, that this consciousness does in factinvolve the recognition of grounds for drinking water and that its contentis “Drink! because this has features that make it thirst-quenching”?

It is indeed true that thirst involves attention to the thirst-quenchingfeatures of water, rather than, say, to its slippery-making features. Myclaim is simply that thirst does not involve consciousness of the thirst-quenching features of water as grounds that justify the felt imperativeto drink. Rather, it involves consciousness of these features as criteriafor fulfilling a more basic imperative of which the imperative “Drinkwater!” is one specification. That more basic imperative is “Quench thisthirst!” I am conscious of water as to-be-drunk because I am consciousof my thirst as to-be-quenched. The thirst-quenching features of waterto which my attention is drawn don’t justify my acting on this morefundamental imperative; they merely show me a way of fulfilling it.

To get the flavor of what I have in mind, consider the way non-human animals are motivated to act. (I am only appealing to intuitions,here, although I think these intuitions can be defended more system-

36. Although imperatives figure centrally in my view, I reject the noncognitivism withwhich they are often associated. As a Kantian, I believe that imperatives, suitably expressedin the form of maxims, can be subject to rational assessment.

37. One might grant that particularly strong or urgent inclinations have this imper-atival structure, while denying that weak ones do. Indeed the phenomenology of theexperience of strong inclination is often different from that of weak inclination. Weakinclinations aren’t as overtly “demanding” as strong ones. It could be argued, then, thatweak inclinations have the structure of proposals, rather than imperatives. On my view,all inclinations have an imperatival structure. I think this not simply because of phenom-enology, but because I believe that all forms of agency, whether rational or nonrational,involve imperatives. To regard an inclination as a proposal is to take a further step, thatof recognizing that the inclination does not have the authority it claims for itself, namely,the authority to dictate my conduct. To see an inclination as a proposal is an achievementrequiring the work of reason. Admittedly I have not yet argued for this in detail. I amgrateful to Sam Scheffler and Josh Cohen for pressing me on this.


atically.)38 How does the world look to a cat? A cat’s world, presumably,is teleologically organized around her needs and interests. The cat seesthis scurrying mouse as to-be-chased, this food in the dish as to-be-eaten,and that big angry dog as to-be-avoided. When she is moved in light ofher teleological consciousness, she is not simply subject to a causaldisposition. The object does not force her to act in the way that pollenin the air might force her to sneeze. Rather, she sees the object as callingupon her to initiate movement in response to it. At least we have tothink of her being motivated in something like this way insofar as weare committed to distinguishing between what she does and what hap-pens to her.

Still, when the cat is motivated in this way, she does not see thedemand to chase the mouse as something requiring justification. In-deed, she does not have the capacity to call her instincts into question(though she can expand their reach through learning).39 So an answerto this justificatory question cannot even implicitly be part of her con-sciousness of the object. Analogously, I claim, when I have an inclinationto reach for that piece of chocolate cake, the inclining part of me is“seeing” the cake as to-be-eaten. It is not also, even implicitly, asking oranswering the question “On what grounds should a creature like medo this sort of thing under these sorts of circumstances?” Just as the catlacks the capacity to call its instincts into question, so the inclining partof me lacks the capacity to call its motivating principles into question.

My contention, then, is that to “have an inclination” is to be awareof a part of me going for something in the way that a nonhuman animalgoes for something. My inclination is the movement of my inner animal,a movement that would count as my action were I wholly a creature ofinstinct.

Now, again, the extreme rationalist might object that there is a ten-sion between saying that the cat acts and saying that the cat does not acton reasons (or that it does not engage in justification). The worry is thatinsofar as we see the cat as an agent, we are committed to describingwhat she does in terms of all the concepts associated with human agency,including those of “justification” and “reason.”40 If we are serious aboutsaying that the cat “acts,” we have to also say that she “freely chooses” toact, and if we are serious about saying that she chooses to act, then wemust claim that she does so on the basis of “reasons” that “justify” her

38. My intuitions about animal agency have been reinforced and elaborated by Chris-tine Korsgaard’s recent work on animal agency (Christine Korsgaard, “Fellow Creatures:Kantian Ethics and Our Duties to Animals,” in The Tanner Lectures on Human Values, vol.25, ed. Grethe B. Peterson [Salt Lake City: Utah University Press, 2005], 79–110, esp. sec.2; this is also available at http://www.tannerlectures.utah.edu).

39. Ibid., sec. 2.40. I am grateful to an anonymous reviewer at Ethics for pressing this objection.


choice. If this is right, then animal agency must have all the basic featuresof human agency, at least in some primitive form. Animal agency mightwell be less clearly articulated than human agency, and so different indegree, but it cannot be different in its basic structure.

My view relies on there being a difference in structure here. Thedifference lies in the fact that the cat cannot gain reflective distance onits own instinctively motivated activity, whereas human beings can. Thecat is conscious, but it is not in this sense self-conscious.41 It can forman evaluative conception of the world in light of which it can guide itselftoward what pleases or interests or satisfies it, but it cannot form anormative conception of itself in relation to the world, in light of whichit can govern itself with respect to these objects. What I mean by this isthat the cat cannot form an ideal of how “a creature like me” ought toconduct itself with respect to “the things that please or interest or satisfyme”; it cannot form a conception of itself that it can uphold or betrayby acting in one way or another.42 Arguably this underlies the deepestdifference in our conceptions of nonhuman and human animals, whichis that nonhuman animals cannot hold themselves or be held account-able for what they do, whereas human animals can.

This might seem to imply only that nonhuman animals cannotengage in moral justification and cannot act on moral reasons, whileleaving open the possibility that they can engage, at least implicitly, inother sorts of justification and act on other sorts of reasons. But I wouldargue that these capacities, the capacity to engage in justification andthe capacity for moral accountability, are indeed continuous. Justifica-tion is a form of address—to another or to oneself-regarded-as-another.Its aim is to show that the action in question is grounded in principlesthat both addressee and addressor could freely accept and to whichthey could hold themselves and one another accountable.43 In order toengage in this kind of address, one arguably has to be self-conscious inthe sense I just described. Since merely conscious creatures lack thecapacity for such self-consciousness, it is sheer anthropomorphism tosee them as engaged, even implicitly, in justification. Merely consciouscreatures are agents who can and must guide themselves through theworld, but this does not imply either that they can or that they mustjustify themselves to one another.

41. Korsgaard uses the “conscious”/”self-conscious” distinction in her account of an-imal agency (Korsgaard, “Fellow Creatures,” sec. 2). The guidance/governance distinctionis mine.

42. I approach the distinction between goal-guided and ideal-governed action froma different direction in my “Three Conceptions of Action in Moral Theory,” Nous 35(2001): 93–117.

43. See Stephen Darwall, The Second-Person Standpoint: Morality, Respect, and Account-ability (Cambridge, MA: Harvard University Press, 2006).


Now this is not an article on the nature of animal agency, and Icannot here fully defend the view that there is this difference in kindbetween animal and human agency. My aim is to say enough in defenseof the view to motivate my conception of inclination. That conceptiondoes not depend directly on any claim about nonhuman animals perse. It simply depends on the claim that there is conceptual room for adistinct kind of agency that does not presuppose justificatory capacitiesand that this kind of agency plausibly characterizes human inclinationas seen from the perspective of one who experiences it. I am going tocall this kind of agency “object-based” in contrast with “principle-based”agency. While the claim that animal agency is object-based provides anintuitive way into my conception of inclination, my fundamental claimis that inclination has the structure of object-based agency. Hence, myposition would not be undermined directly if it turned out that animalagency does not exemplify object-based agency. Still, it would not beimplausible to think that our capacity to be inclined is evolutionarilyrelated to agential capacities we share with nonhuman animals. In ad-dition, the association of inclination with animal agency lends supportto the everyday idea that to act out of inclination is to “act like ananimal.”

The examples I have offered so far have involved inclinations stem-ming rather directly from our biological nature, ones we share withmany nonhuman animals. But I do not want to suggest that the onlypossible objects of human inclination are those that are also objects ofanimal action. The point is that the mode of agency involved is thesame; the objects can, of course, be different. Human beings can beinclined to do a whole range of things that nonhuman animals simplycould not be motivated to do. Let me illustrate with a different example.Suppose I want to check my e-mail. Part of this motive might be practicalrather than pathological. It might involve the thought that I ought tocheck my e-mail because I am involved in an important negotiation andI risk losing out if I don’t keep in touch in a timely manner. I will setthat practical motive aside. The pathological part of the motive, theinclination, is simply an urge to check my mail. On my view, this motiveinvolves my consciousness of something like an imperative, “Check e-mail!” Moreover, in experiencing the normative pull of this imperative,I attend to the features of e-mail that make it appealing to me. Supposethat, in my case, the appealing feature of e-mail is not its usefulness asa professional tool but rather its effectiveness at keeping me in contactwith family and friends. If this is the feature of e-mail that I attend toin having the inclination to check it, then my underlying felt imperativemay be something like “Stay connected with friends and family!” Myinclination to check e-mail is an urge to do so as a way of staying con-nected with friends and family. Now, in order to be influenced in this


way, do I need to be conceived of as offering myself, even implicitly, ajustification for staying connected with friends and family? No. The feltneed to stay in touch, like the felt need to quench my thirst, is guidingthe inclining part of me in the way that instinct guides an animal.44

A similar story can be told about the range of inclinations thatshape my personal tastes and interests. Suppose I feel like listening tomy current favorite piece of music. Again, assuming this motive is notthe result of prior practical reasoning (e.g., I ought to listen to this songbecause doing so would make my friend happy), it simply involves myconsciousness of this piece as to-be-listened-to. Indeed, I might say thatI want to drink this piece of music in through my ears. Insofar as thisstatement makes sense, it is because the inclination involved shares thesame structure with thirst. My inclination to listen to the music is nomore laden with justificatory thought than is my inclination to drinkwater when thirsty. This is not to take a position on the question whetherjudgments of taste can in principle be justified. The point is simply thatinclinations based on taste do not have to involve justificatory thoughtin order to have motivational influence.

I have claimed that having an inclination involves having and beingmotivationally responsive to an imperatival conception of an object andthat this does not involve the further thought of a justification for re-sponding as the imperative directs. This amounts to a rejection of Scan-lon’s view of desire because, on that view, desire involves taking a con-sideration as a reason to act. But does it amount to a rejection of Quinn’sview, which claimed that the desire necessarily involves some conceptionof the object as good? Recall that, according to Quinn, my desire hasto involve my evaluation of an object as good in a broad sense (asworthwhile, desirable, attractive, pleasant, etc.). If it didn’t, it wouldn’tnecessarily provide me with a perspective from which to see the actionon the desire as having any point. Now Quinn might object that, onmy view, inclination fails to rationalize the response it motivates because,in being inclined, I simply find myself responding to imperatives I can-not justify. Why don’t these imperatives seem arbitrary from my pointof view, and why don’t my motivational responses to them seem pointlessto me?45

A way of getting at the issue here is to ask what my view would sayabout Quinn’s example of the disposition to turn on radios. Suppose Ifind myself inclined to turn on every radio I see. On my theory, thismeans that I see every radio that is off as to-be-turned-on. I have an

44. I don’t mean to be suggesting here that this need is specifically human. Manykinds of nonhuman animals need and thrive on affective bonds with other individuals.

45. I am grateful to both Agnes Callard and an anonymous reviewer at Ethics forurging me to address this objection.


imperatival conception of such radios, and at least part of me is moti-vationally responsive to this conception. Do I need to have the furtherthought that turning on radios is in some way good? If not, then howis the motivation I have described any more intelligible to me than, say,my disposition to squint in bright sunlight?

When I squint in bright sunlight, I do not do so in response to animperatival conception of anything. I do not see the sun as to-be-squinted-at or sunlight as to-be-squinted-in. I do not have an inclinationto squint and then act on that inclination. Rather, I squint by reflex.So, on my theory, my inclination to turn on radios is not assimilable toa reflex. But it might still seem objectionably arbitrary. Suppose all theobjects in my world have little labels on them telling me what I oughtto do in response to them. My bed says “to-be-slept-in,” and my chairsays “to-be-sat-upon.” All the radios say “to-be-turned-on.” If my accountis interpreted as claiming that having an inclination is like followingthe instructions on these labels for no reason, then it will run afoul ofQuinn’s worry. The motivation will make no sense to me from the inside.

To rule out this caricature, I have to flesh out the details of mytheory in a bit more detail. I have already said that, on my view, incli-nation involves seeing an object in an imperatival way. This is differentfrom seeing the object and then seeing an imperative attached to theobject. When I am inclined, my attention is necessarily drawn to certainfeatures of the object that appear to me as practically salient. If I havean inclination to turn on radios, this must involve seeing radios as to-be-turned-on in virtue of certain features of radios and of the action.Part of what makes the disposition that Quinn describes seem unintel-ligible is that it does not specify any features as practically salient fromthe perspective of one who has this disposition. On my view, it must bepossible to specify such features if the motive is to count as an incli-nation. Let’s suppose that seeing a radio in the off position simplygenerates a brute, unanalyzable feeling of psychological discomfort inme; I experience turned-off radios as something like mental itches Ihave to scratch. On the theory I have been defending, this is sufficientfor the impulse to count as an inclination. It counts as an inclinationbecause it is the kind of impulse that can exert an influence on my will.It is no doubt less intelligible to us than other inclinations because itis uncommon and because there is no familiar physiological mechanismunderlying it, but that does not disqualify it as an inclination.

This is not inconsistent with Quinn’s view. To see an off radio as amental itch is to see it in terms of pain or discomfort and, accordingto Quinn, pain and pleasure are primitive evaluative notions, primitivenotions of bad and good. I agree with this view of pleasure and pain,though I cannot argue the point here. Now, I am willing to grant that,when I am inclined, I necessarily see a certain action as to-be-done in


virtue of features that make doing it look good to me. Some version ofthe “guise of the good” thesis does characterize object-based agency.But, again, this does not mean that I see the goodness of those featuresas justifying my doing the action. What it means is that looking-good-to-me functions as the basic criteria any action must fill in order for itto appear to me as to-be-done. In other words, each of my inclinationsmanifests my inclining self ’s responsiveness to a basic imperative to seekmy apparent good and to shun my apparent bad. This is consistent withthe analogy to animal action, because creatures of instinct character-istically act in light of what their sense of their weal and woe. If myinclining self weren’t in general sensitive to something I see as my wealand woe, no actions would spontaneously appear to me as to-be-done.But this does not imply either that the primitive concept of goodnessis prior to the primitive concept of practical necessity or vice versa. Thetwo stand and fall together. My inclining self cannot be conscious ofanything as to-be-sought without conceiving of it as good, and it cannotbe conscious of anything as good without conceiving of it as to-be-sought.

IV. IMPLICATIONS OF THE THEORY

The view of inclination as stemming from our capacity for nonrationalagency can account for the features of inclination that caused problemsfor both extreme anti-rationalists and extreme rationalists. Recall thatthe main problem for the extreme anti-rationalist was to explain thedistinctive role inclination plays in deliberation. In ordinary delibera-tion, inclinations do not present themselves as features of our circum-stance, as on par with the weather. Rather, they present themselves asdemands or proposals or suggestions about what to do given our cir-cumstances. On my view, having an inclination involves having an im-peratival conception of an object. As such, the consciousness involvedin inclination purports to direct us, to tell us what to do. This is why itis something we can act on, satisfy, and side with.

The main challenge for the extreme rationalist was to explain thevarious ways in which we are passive with respect to our inclinations.One aspect of this challenge was to explain how inclination can conflictwith reason. Now notice that conflict requires both similarity and dif-ference. Extreme anti-rationalism cannot account for conflict betweeninclination and reason because it assimilates inclination to a happeningrather than a mode of government. Extreme rationalism assimilatesinclination to a mode of government indistinguishable in kind fromreasoning, and so it cannot account for the duality necessary for genuineconflict. On my view, inclination and reason are similar to and differentfrom one another in ways that make conflict possible. They are both inthe business of governing action, and as such they can issue conflictingclaims about what to do. But they are different modes of government,


which presuppose different agential capacities. Our inclination ad-dresses us as nonrational animals subject to the dictatorial authority ofinstinct. Our reason addresses us as free beings capable of autonomy.In this respect, they are two different motivational sources.

A related challenge is to explain how reason and inclination canconflict while still accounting for the recalcitrance of inclination. Thesefeatures of the relation between inclination and reason are in tensionwith one another. To the extent that reason and inclination can comeinto direct conflict, they must be in the same business. But to the extentthat inclination, by its very nature, is not directly responsive to volition,the two have to be operating according to different principles. I notedearlier that one of the intuitions that drives anti-rationalism is the in-tuition that I cannot change my inclinations simply by changing mymind. Now the extreme rationalist might agree with this, while sayingthat, to the extent that I can’t change my inclinations by changing mymind, I am simply being obtuse. For if my inclinations are just my hasty,unreliable acts of judgment, then I should be able to change them simplyby making more careful, reliable judgments. So long as I am rational,I should in principle be able to cultivate my inclinations through strictlyrational means.

Part of why the perceptual analogy is compelling is that it puts thelie to this way of thinking. I may be perfectly rational and still see thestick in the water as bent even though I judge it to be straight. I maybe perfectly rational and still feel a rush of fear when riding the rollercoaster, even though I judge that I am completely safe. My accountprovides an explanation for why the latter situation is both one of mo-tivational conflict and one of recalcitrance for which I am not at fault.The part of me that experiences the fear (which, for my purposes, wecan construe as an inclination to flee the situation) is what I have calledmy “inner animal.” The part of me that judges that I am safe (and sotells me I can stay put) is what I will call my “outer human.” The situationcounts as a motivational conflict because my inner animal tells me toflee while my outer human tells me to stay put. It also illustrates therecalcitrance of inclination, because my inner animal’s way of seeingand responding to the situation does not change simply because myouter human has judged the situation differently. But this recalcitrancedoes not stem from obtuseness on my part. My inner animal simply isnot capable of seeing the situation in exactly the same terms that myouter human sees it. Although it sees the situation as to-be-fled, it doesnot see it as to-be-fled-for-a-reason, and so it cannot be corrected simplyby being shown a stronger reason to stay put. Now, again, in saying thatmy inner animal does not see the situation as to-be-fled-for-a-reason, Iam not denying that it sees the situation as to-be-fled in virtue of specificfeatures of the roller coaster ride, in this case its life-threatening features


rather than, say, its noisy features. But this is just to say that my inneranimal is responding to a more fundamental imperative, “Preserve mylife!” in light of which it sees the situation.

At this point, the extreme rationalist might object. When I’m stand-ing in line for the roller coaster, I may well try to talk myself out of thefear I’m feeling. But if inclinations are simply not responsive to reason,why would doing this make any more sense than trying to talk myselfout of indigestion? Even Aristotle, who was certainly not an extremerationalist, held that the irrational part of the soul “listens to” and “isin some sense persuaded by” the rational part.46 But the neo-Aristotelianpicture that I have defended thus far does not explain how this is pos-sible. How do we account for the fact that talking to oneself can be aneffective way to try to manage one’s inclinations if the inclining part ofthe soul is by definition not capable of engaging in rational dialogueabout what to do?

My reply has two parts. First, I assume that my inner animal’s con-sciousness is capable of making associative connections. In the rollercoaster case, my inner animal’s tendency to see roller coasters as life-threatening is not the product of its having gathered and assessed evi-dence in a rational way. It has, by nonrational processes, come to seeroller coasters as fragile, rickety, and unreliable. Now, my rational partcan, by assessing the evidence, come to judge that roller coasters arenot, in fact, unreliable. But for this knowledge to shape my inclinations,I need something more than rationality. I need to forge new associativeconnections, perhaps through the exercise of my imagination. I haveto come to see roller coasters as solid, supportive, and trustworthy. Some-times when I try to talk myself out of my fear, what I am doing isattempting imaginatively to forge these new associative connections.

Second, when I am in the roller coaster line, what looks like anattempt to talk myself out of having my fear is often in fact an attemptto talk myself out of acting on my fear. I may be trying to reason withmy rational part, the part that makes choices in the face of inclination.47

Reminding myself of the evidence for thinking that roller coasters aresafe can indeed help to keep my will from siding with my inclination.Moreover, by acting as if I am not afraid of roller coasters, I give myselfthe chance to have experiences that can help me generate and reinforcethe associative connection I need to support my rationality, the con-nection between roller coasters and stability.

My account can thus explain both the sense in which inclinationsare, by their very nature, recalcitrant to reason and the sense in whichthey are responsive to certain modes of conscious control that involve

46. Aristotle, Nicomachean Ethics, 1102b13–1103a.47. Thanks to Jason Slavick for helping me formulate this part of the reply.


talking to ourselves. The final challenge is to explain why we are notresponsible for having our inclinations in the same way that we areresponsible for acting on them. The answer should by now be clear.While my inclinations are not simply happenings external to me, neitherdo they issue from the part of me that can be held directly accountablefor what it does. The capacity in virtue of which I have inclinations lacksthe freedom necessary for accountability. It is agential, but it is governedby laws it cannot question. Hence, my inner animal does not speak forme in the deep sense that my outer human does.

V. CONCLUSION

I have developed and defended a view according to which our capacityfor inclination is both agential and nonrational. In developing this con-ception, I have tried to learn from the insights of both anti-rationalismand rationalism, without taking either to an extreme. The anti-rationalistinsight is that inclinations stem from a motivational source distinct fromreason or will. The rationalist insight is that inclination engages ouragential capacities by motivating us on the basis of our evaluative out-looks. If inclination is the exercise of a capacity for nonrational agency,both of these insights can be true. On my view, our inclinations stemfrom a capacity for object-based agency, the form of agency that I suggestcharacterizes creatures of instinct. Object-based agency is consciouslyguided and so is not simply a way of being caused. But it is not rationalin that it does not engage the capacity to raise and answer questions ofjustification. To be influenced by inclination, I conclude, is to be awareof a part of me—my inner animal, so to speak—going for somethingin the way that would count as my action were this part the whole ofme.

Let me close by noting the limits of this argument. I have not yetaddressed the question of how inclination and reason work together togenerate action. Importantly, I have not yet said anything about whetherinclination alone is sufficient to determine human action or whether itcan only do so with the cooperation of the will. Nor have I attemptedto explain how inclination provides input to deliberation and how rea-son takes up this input. And although I have attempted to provide anexplanation of motivational conflicts where inclination is recalcitrant toreason, I have not attempted to give a theory of akratic action. My theorywill have implications for all of these questions, but I assume that work-ing them out is not trivial. Indeed, there is a tendency in philosophyof action to conflate claims about what is involved in having inclinationswith claims about what is involved in acting on them. I have proceededon the assumption that it is wise to address the first of these issues beforeturning to the second.

the nature of inclination* tamar schapiro

Documents